Impact of Insulin Receptor Substrate-1 Genotypeson Platelet Reactivity and Cardiovascular Outcomesin Patients With Type 2 Diabetes Mellitusand Coronary Artery Disease
Dominick J. Angiolillo, MD, PHD,* Esther Bernardo, BSC,† Martina Zanoni, PHD,‡David Vivas, MD, PHD,† Piera Capranzano, MD,* Giovanni Malerba, PHD,‡Davide Capodanno, MD,* Paola Prandini, PHD,‡ Alessandra Pasquali, PHD,‡Elisabetta Trabetti, PHD,‡ Manel Sabaté, MD, PHD,† Pilar Jimenez-Quevedo, MD, PHD,†Jose L. Ferreiro, MD,* Masafumi Ueno, MD,* Theodore A. Bass, MD,* Pier Franco Pignatti, MD,‡Antonio Fernandez-Ortiz, MD, PHD,† Carlos Macaya, MD, PHD†
Jacksonville, Florida; Madrid, Spain; and Verona, Italy
Objectives The aim of this study was to assess the association between genetic variants of the insulin receptor substrate(IRS)-1 gene, platelet function, and long-term outcomes in patients with type 2 diabetes mellitus (DM) and sta-ble coronary artery disease while on aspirin and clopidogrel therapy.
Background The effects of pharmacogenetic determinants on platelet function and cardiovascular outcomes in type DM pa-tients are unknown.
Methods The association between IRS-1 genetic variants, platelet function, and the risk of major adverse cardiac events(MACE) at 2 years was assessed in 187 patients with type 2 DM and stable coronary artery disease on mainte-nance aspirin and clopidogrel therapy.
Results Seven tag single nucleotide polymorphisms were selected. Individuals with high platelet reactivity were more frequentamong carriers of the C allele (GC and CC genotypes; approximately 20% of population) of the rs956115 marker(44.4% vs. 20.5%; odds ratio: 3.1, 95% confidence interval [CI]: 1.44 to 6.67; p � 0.006). These patients were athigher risk of MACE (28.0% vs. 10.9%; hazard ratio: 2.90, 95% CI: 1.38 to 6.11; p � 0.005). The C allele carriers ofthe rs956115 marker were more commonly associated with a hyperreactive platelet phenotype. This was confirmedin an external validation cohort of patients with type 2 DM but not in an external validation cohort of patients withoutDM. Carriers of the C allele of the rs956115 marker also had a significantly higher risk of MACE compared with non-carriers (30.6% vs. 11.4%; hazard ratio: 2.88, 95% CI: 1.35 to 6.14; p � 0.006).
Published by Elsevier Inc. doi:10.1016/j.jacc.2011.02.040
Dual antiplatelet therapy with aspirin and clopidogrel is therecommended treatment for patients with acute coronarysyndromes (ACS) and in those undergoing percutaneous
From the *Division of Cardiology, University of Florida College of Medicine,Jacksonville, Florida; †Cardiovascular Institute, San Carlos University Hospi-tal, Madrid, Spain; and the ‡Department of Life and Reproduction Sciences,University of Verona, Verona, Italy. This work has been partially supported by agrant from the Spanish Ministry of Health, the Italian Ministry of University andResearch, and the Veneto Region Sanitary Research Project. Dr. Angiolillo wassupported in part by the Marie Curie Individual Fellowship Grant and hasreceived honoraria for lectures from Bristol-Myers Squibb, Sanofi-Aventis, Eli
Lilly Company, and Daiichi Sankyo, Inc.; consulting fees from Bristol-MyersSquibb, Sanofi-Aventis, Eli Lilly Company, Daiichi Sankyo, Inc., The Medicines
coronary interventions (PCI) (1,2). Numerous investiga-tions have shown a broad variability in interindividualresponse to antiplatelet therapy, and patients with high
Company, Portola, Novartis, Medicure, Accumetrics, Arena Pharmaceuticals,AstraZeneca, and Merck; and research grants from Bristol-Myers Squibb,Sanofi-Aventis, GlaxoSmithKline, Otsuka, Eli Lilly Company, Daiichi Sankyo,Inc., The Medicines Company, Portola, Accumetrics, Schering-Plough, Astra-Zeneca, and Eisai. Dr. Bass has received honoraria for lectures from Eli LillyCompany and Daiichi Sankyo, Inc.; consulting fees from Eli Lilly Company andDaiichi Sankyo, Inc.; and research grants from Baxter. All other authors havereported that they have no relationships to disclose.
Manuscript received December 13, 2010; revised manuscript received January 27,2011, accepted February 22, 2011.
ts(lmdwstfIlcawv
M
SamidspAcetccdpcP
csosotpaaadagmdmm
31JACC Vol. 58, No. 1, 2011 Angiolillo et al.June 28, 2011:30–9 Gentics, Diabetes, and Outcomes
on-treatment platelet reactivity (HPR) have an increasedrisk of ischemic events (3,4). Patients with diabetes mellitus(DM) have a greater prevalence of HPR compared withnon-DM, which might explain their overall enhanced riskof developing atherothrombotic complications (5–8). How-ever, heterogeneous antiplatelet drug effects might also beobserved among patients with DM, and platelet functionprofiling even within this high-risk cohort identifies subjectsat a greater risk of recurrent ischemic events (9). Themechanisms leading to variable antiplatelet drug responseprofiles in patients with DM are not fully elucidated.Although there is growing evidence that single nucleotidepolymorphisms (SNPs) might modulate antiplatelet drugeffects (10), whether these might explain the heterogeneityin response profiles and clinical outcomes selectively inpatients with DM remain unexplored.
Human platelets are targets of insulin effects that aremediated by the insulin receptor substrate (IRS)-1 (8,11). Inhealthy volunteers, insulin interferes with calcium increasesinduced by adenosine diphosphate (ADP)-P2Y1 contacthrough Gi activity and, thereby, with P2Y12-mediateduppression of cyclic adenosine monophosphate (cAMP)11). However, platelets from patients with type 2 DM haveost responsiveness to insulin, leading to increased P2Y12-
ediated suppression of cAMP and decreased antiplateletrug effects (12). Importantly, studies performed in subjectsithout DM or in a pre-DM status have shown that gene
equence variations of IRS-1 are associated with the func-ional activity of this receptor (13) as well as being a riskactor for coronary artery disease (14). However, whetherRS-1 genotypes are associated with variations in antiplate-et drug response profiles and whether these might impactlinical outcomes in patients with DM is unknown. Toddress this issue we evaluated whether IRS-1 genotypesere associated with platelet function profiles and cardio-ascular outcomes.
ethods
tudy population. Blood samples for platelet functionnalyses and genotyping were collected from a total of 208edically treated (with oral hypoglycemic agents and/or
nsulin) patients with type 2 DM and stable coronary arteryisease from November 2003 to March 2007. To avoidtratification of the sample due to ethnicity, only Caucasianatients homogeneous for ethnic background were included.ll patients (primary cohort as well as 2 external validation
ohorts) were from the central regions of Spain. To beligible, patients with type 2 DM (�18 years of age) neededo have undergone PCI and been receiving aspirin andlopidogrel therapy for 6 to 9 months in the absence ofardiovascular events during this period. Type 2 DM wasefined according to the World Health Organization Re-ort (15). All patients were recruited from the outpatientlinic of our hospital as part of their routine follow-up after
CI. Aspirin (100 mg/day) was used indefinitely, and
lopidogrel (75 mg/day) was pre-cribed for 12 months after cor-nary revascularization. Bloodampling was not performed if 1f the following exclusionary cri-eria was present: 1) use of anti-latelet agents other than aspirinnd clopidogrel; 2) use of oralnticoagulants; 3) occurrence ofn acute cardiovascular eventuring the interval between PCInd blood sampling; 4) impairedlucose tolerance without phar-acologic treatment, gestational
creatinine levels �2.5 mg/dl; or8) hepatic enzymes (alanine ami-notransferase or aspartate amino-transferase) twice the upper nor-mal limit.
Patients meeting study eligi-bility criteria were followed for24 months, and clinical eventswere recorded. Follow-up wasperformed by means of telephonecontacts every 6 months and clinic visits on a yearly basis.Patients with nonvaluable pharmacodynamic assessmentswere excluded from the final analysis. The primary outcomemeasure was a composite of cardiovascular death, ACSleading to hospital stay, and nonfatal stroke. Such majoradverse cardiovascular events (MACE) were defined ac-cording to definitions proposed by the American College ofCardiology (see the Online Appendix for complete descrip-tion) (16). The treating physicians and investigators whoadjudicated the clinical endpoints were blinded to the resultsof the pharmacodynamic and genotype assessments.
After our initial investigation to define the prevalence andfunctional impact of IRS-1 genotypes in our main studycohort, validation assessments were performed to replicatethe pharmacodynamic findings associated with the carrierstatus of the C allele of the rs956115 marker, whichemerged from the main cohort to be associated with ahyperreactive platelet phenotype. In particular, a separateexternal cohort of patients with type 2 DM (n � 52)undergoing elective PCI was identified to confirm thepharmacodynamic impact of this marker in the acute phaseof clopidogrel therapy. All patients were taking aspirintherapy and received a 600-mg loading dose of clopidogrelat the time of intervention; pharmacodynamic assessmentswere performed at hospital discharge. Furthermore, todetermine whether or not the pharmacogenetic findingswere specific to patients with DM, a pharmacodynamicassessment was also extended to a cohort of patients without
Abbreviationsand Acronyms
ACS � acute coronarysyndrome
ADP � adenosinediphosphate
CI � confidence interval
CYP � cytochrome P450
DM � diabetes mellitus
HbA1C � hemoglobin A1C
HPR � high plateletreactivity
HR � hazard ratio
IRS � insulin receptorsubstrate
LD � linkage disequilibrium
MACE � major adversecardiovascular event(s)
OR � odds ratio
PCI � percutaneouscoronary intervention
ROC � receiver-operatorcharacteristic
SNP � single nucleotidepolymorphism
DM (n � 90). Similarly, to study
subjects from the main
ludpsGdllwAUEPqBpSpiotSAsfcm(GHeoSfttwW
w(1w
ah
32 Angiolillo et al. JACC Vol. 58, No. 1, 2011Gentics, Diabetes, and Outcomes June 28, 2011:30–9
cohort, these patients were in their steady state phase of thesame doses of aspirin and clopidogrel therapy. A flowdiagram of the main study cohort and separate externalvalidation cohorts is provided in the Online Appendix(Online Fig. 1).
The study complied with the Declaration of Helsinki andwas approved by the Ethical Committee of the San CarlosUniversity Hospital, and all patients gave their informedconsent.Pharmacodynamic assessments. Pharmacodynamic effectswere assessed with light transmittance aggregometry ac-cording to standard protocols, as previously described (seethe Online Appendix for complete description) (5,6,9).Maximum platelet aggregation was measured after stimuliwith ADP (20 �mol/l) to assess purinergic mediated plate-et function. High platelet reactivity was defined as thepper quartile of ADP-induced aggregation, as previouslyescribed (9,17,18). To define nonpurinergic mediatedlatelet function, platelet aggregation after collagen (6 �g/ml)timuli was also performed (6,9).
enotyping and haplotype association analyses. Genomiceoxyribonucleic acid was extracted from peripheral-blood
eucocytes with standard salting-out procedures. The se-ection of the tag SNPs of IRS-1 gene was performedith GEVALT 2.0 software (GEnotype Visualization andLgorithmic Tool) (19–21). The SNP genotype data fortah residents with ancestry from northern and westernurope (CEU) population were downloaded from HapMaproject Browser, submitting a 100-kilobase pair region as auery (chr2: 227,290,450..227,390,449; release April 2007).ecause rs1801278 has been extensively investigated in theublished data, this was force-included in the list of tagNPs identified (13,14). The 7 tag SNPs gave an estimatedrediction value of 97.6% for the IRS-1 genomic regionnvestigated. Genotyping was performed with FRET (Flu-rescent Resonance Energy Transfer) Probes technology onhe LightCycler 2.0 instrument (Roche Diagnostics, Basel,witzerland) and TaqMan SNP Genotyping assays on andpplied Biosystems StepOnePlus instrument (Applied Bio-
ystems, Foster City, California) (see the Online Appendixor complete description). Lewontin’s D= and the square oforrelation coefficient r between 2 markers were used as aeasure of linkage disequilibrium (LD) of all marker pairs
22). The LD haplotype block structure was identified withEVALT 2.0 according to the gerbil algorithm (20).aplotype association analyses were performed on the
ntire haplotype region (i.e., including the 7 tag SNPs) andn subregions defined by LD blocks (23).tatistical analysis. Continuous variables were analyzed
or a normal distribution with the Kolmogorov-Smirnovest and presented as mean � SD or median and interquar-ile range, as appropriate. Normally distributed variablesere analyzed with Student t tests, whereas the Mann-hitney U test was used for comparisons of non-normally
distributed variables. Categorical variables are presented as
frequencies and percentages and were compared with the
use of chi-square test or Fisher exact test where appropriate.Receiver-operator characteristic (ROC) analyses were per-formed for an exploratory evaluation of the optimal cutoffvalue of ADP- and/or collagen-induced platelet aggregationfor predicting MACE in our study population (9,24).Spearman’s rank correlation was used to examine the cor-relation between profiles of platelet reactivity and glycemiccontrol, defined by hemoglobin A1C (HbA1C) levels.Interaction among genotype, insulin resistance (defined byhomeostatic model assessment; see Online Appendix fordescription), and HPR was also determined. Rates of theprimary endpoint are expressed as Kaplan-Meier estimatesat 24 months and compared with log rank testing. Univari-able and multivariable Cox proportional hazards regressionmodels were used to assess unadjusted and adjusted risk ofthe combined cardiovascular endpoint associated with HPRand genotype. Demographic, clinical, and laboratory vari-ables provided in Table 1 were entered in the Cox model forthe multivariable analysis, and those that were not signifi-
Baseline Demographic Data and ClinicalCharacteristics According to Post-TreatmentPlatelet Reactivity StatusTable 1
Baseline Demographic Data and ClinicalCharacteristics According to Post-TreatmentPlatelet Reactivity Status
VariableHPR
(n � 47)No HPR
(n � 140) p Value
Age (yrs) 64 � 10 67 � 10 0.15
Male 31 (66) 101 (72) 0.54
Risk factors/past medical history
Insulin-treated diabetes 20 (43) 41 (29) 0.13
HbA1C 7.4 � 1.3 7.1 � 1.1 0.18
Smoking 9 (19) 19 (14) 0.49
Hyperlipidemia 30 (64) 106 (76) 0.16
Hypertension 31 (66) 91 (65) 0.95
Body mass index (kg/m2) 28.3 � 4.1 28.8 � 3.7 0.44
Obesity (body mass index �30 kg/m2) 19 (40) 52 (37) 0.82
Values are n (%) or mean � SD. *Aspirin and clopidogrel were used in all (100%) patients. Thereere no differences in length of clopidogrel therapy between high on-treatment platelet reactivity
HPR) and non-HPR groups (total duration: 12 months; post platelet function assessment: 4.78 �
.49 months vs. 4.63 � 1.50 months, p � 0.53). Among noninsulin-treated diabetic subjects, thereere no differences in type of oral hypoglycemic agents used between HPR and non-HPR groups.ACE � angiotensin-converting enzyme; ARB � angiotensin II receptor blockers; CABG � coronary
cant at p � 0.10 were removed by a backward stepwise
BseaapOtatmhstrt((ft
acttthcwhT(ehtwA
((aeat�tno
R
Ctsbe(aropbsqoH�tawswvgat4AwtI(r9dfrptasH
33JACC Vol. 58, No. 1, 2011 Angiolillo et al.June 28, 2011:30–9 Gentics, Diabetes, and Outcomes
elimination. With this method, chronic renal insufficiencywas identified as the only significant predictor associatedwith the primary endpoint. The HPR and genotype wereadded as independent categorical variables in the model,including chronic renal insufficiency as variable for statisticaladjustment. The assumption of proportional hazard waschecked with time-dependent covariates and was found tobe reasonable. First order interactions were evaluated. Haz-ard ratios (HRs) and 95% confidence intervals (CIs) werecalculated. Odds ratios (ORs) are provided for the associa-tion between ADP and/or collagen-induced aggregationabove the upper quartile with the genotype. A p value �0.05was considered statistically significant for all the testsmentioned in the preceding text. Statistical analysis wasperformed with SPSS software (version 14.0, SPSS, Inc.,Chicago, Illinois).
Hardy-Weinberg equilibrium was evaluated for each tagSNP, and markers were rejected if they violated Hardy-Weinberg equilibrium with a threshold of p � 0.01.
onferroni correction was applied to adjust the nominalignificance level of the association test of HPR status withach of the 7 tag SNPs (p � 0.05/7 � 0.007). A multivari-ble logistic regression analysis, including the genotypelong with all the covariates that might impact the degree oflatelet aggregation, was performed to assess the adjustedR for ADP and/or collagen-induced aggregation above
he upper quartile, associated with the genotype. The ADPnd/or collagen-induced aggregation above the upper quar-ile were treated as a dependent variable, and age, sex, bodyass index, diabetes status (insulin- or noninsulin-treated),
yperlipidemia, hypertension, smoking, HbA1C, renal in-ufficiency, and concomitant medications were included intohe statistical model as covariates. All probability valueseported are 2-sided, and a value of p � 0.05 was consideredo be significant. The SNPs showing significant associationsp � 0.05) were then tested for recessive or dominant modeli.e., grouping the heterozygotes together with homozygotesor the major allele or for the minor allele, according withhe model).
Generalized linear models were used to assess haplotypessociations while adjusting for the effects of nongeneticofactors. The null hypothesis of no haplotype effects wasested by standard methods that compare the deviances ofhe model including or not including genetic data (globalest). The significance of the effect of each individualaplotype was also tested (individual haplotype test). Asso-iation of individual haplotypes was considered significanthen both p values of global and at least 1 of the individualaplotype tests were below a threshold value (p � 0.05).he effect of each haplotype was assumed to be additive
i.e., a linear increase in log-odds). Haplotypes whosestimated frequency was �0.01 were grouped into “rare”aplotypes and then treated as a single haplotype. Haplo-ype association analyses were computed with the R soft-are (R Foundation for Statistical Computing, Vienna,
ustria) and the library “haplo.stats” (25). (
We estimated that this study has enough statistical powerbeta � 0.80) to detect an association between HPRdefined as ADP-induced platelet aggregation �64%) andn SNP at the significance level of 0.05 under the hypoth-sis that the risk allele has a frequency of 10% and is inbsolute LD (D= � 1) with the causative variant and thathe OR of the carrier of the risk allele versus noncarrier is
2.8. The number of patients determined to be included inhe validation samples was approximately one-third of theumber of patients included in the main cohort, as previ-usly established (26).
esults
haracteristics of the patients and platelet reactivity. Ofhe 208 patients enrolled in the main cohort of the presenttudy, pharmacodynamic and genotype assessments wereoth available in 187 (89.9%), who were therefore consid-red for the present analysis. The remaining 21 patients10.1%) were excluded, due to inability to measure plateletggregation for reasons including hemolysis, low platelet-ich-plasma platelet counts (�150,000/�l), and instabilityf tracings. In the overall study population, ADP-inducedlatelet aggregation was 55 � 15% and followed a normalell-shaped distribution indicative of a heterogeneous re-ponse profile. The ADP-induced platelet aggregationuartile cut points for the 25th, 50th, and 75th percentilesf the study population were 45.0%, 55.0%, and 64.0%. ThePR was defined as ADP-induced platelet aggregation64%. Baseline demographic data and clinical characteris-
ics of patients with (n � 47) and without (n � 140) HPRre provided in Table 1. Insulin-treated diabetic subjectsere more frequent in the HPR group, although no
tatistically significant differences were found. Also, thereere no significant differences between groups for all otherariables. In the overall population, collagen-induced aggre-ation was 45 � 19%. Quartile cut points for the 25th, 50th,nd 75th percentiles were 33.0%, 46.0%, and 59.0%, respec-ively. Collagen-induced aggregation was 58 � 15% versus1 � 18% in patients with and without HPR defined withDP stimuli, respectively (p � 0.0001). Among patientsith HPR, 55.3% had collagen-induced aggregation above
he 75th percentile.RS-1 genotypes and platelet reactivity. Seven tag SNPsrs11683087, rs2251692, rs1801278, rs1801123, rs6725330,s1896832, rs956115) with an estimated prediction value of7.6% were selected. None of these 7 SNPs showedeviation from Hardy-Weinberg equilibrium, and theirrequencies were similar to those reported in the Utahesidents with ancestry from northern and western Europeopulation. Table 2 summarizes marker information andhe observed genotype frequencies for the 7 tag SNPsssessed. Of the 7 tag SNPs, only the rs956115 markerhowed a significant association with HPR. Individuals withPR were more frequent among carriers of the C allele
GC and CC genotypes) of the rs956115 marker (44.4% vs.
wBTs0mibiHtrPa
D1titn
atPbgmpara1awtIhTpstsg(rTsLbTL0i0ntg1
34 Angiolillo et al. JACC Vol. 58, No. 1, 2011Gentics, Diabetes, and Outcomes June 28, 2011:30–9
20.5%; OR 3.1, 95% CI: 1.44 to 6.67; p � 0.006) (Fig. 1A),hich remained statistically significant when applying theonferroni correction (p � 0.007 required for significance).his association was also confirmed in an adjusted regres-
ion analysis (adjusted OR: 3.56, 95% CI: 1.51 to 8.38; p �.004). The prevalence of C allele carriers of the rs956115arker increased across quartile distribution of ADP-
nduced aggregation (Fig. 1B). There was no correlationetween ADP- (r � 0.092, p � 0.210) and collagen-nduced (r � 0.094, p � 0.200) platelet reactivity and
bA1C levels. Furthermore, there was no interaction be-ween C carrier status of the rs956115 marker and insulinesistance in determining HPR (p for interaction � 0.419).roportions of patients with HPR after collagen and ADP-nd collagen-induced stimuli are reported in Table 3.
In the external validation cohort of patients with type 2M, C allele carriers of the rs956115 marker represented
9.2% of the patient population. These patients werereated with a 600-mg loading dose of clopidogrel, resultingn more suppressed platelet reactivity compared with pa-ients from the main cohort who were receiving mainte-ance therapy. The ADP- and collagen-induced platelet
Tag SNPs From the IRS-1 GeneRegion and Genotype FrequencyTable 2 Tag SNPs From the IRS-1 GeneRegion and Genotype Frequency
Values are n (%), unless otherwise indicated. Position � position (nucleotides) on chromosome 2according to single nucleotide polymorphism database (dbSNP) (Map to Genome Build:36.3).
IRS � insulin receptor substrate; MM � homozygote for major (most frequent) allele M; Mm �
heterozygote; mm � homozygote for minor allele m.
Figure 1 Prevalence of High On-Treatment Platelet Reactivity A
Carriers of the C allele (GC and GG genotypes; solid bars) of the rs956115 genotyon-treatment platelet reactivity compared with noncarriers (GG genotype; open bartion of platelet reactivity (B).
ggregation values after clopidogrel loading dose adminis-ration were 30.7 � 29.5% and 25.9 � 24.2%, respectively.roportions of patients with HPR after ADP, collagen, andoth ADP- and collagen-induced stimuli were significantlyreater among carriers of the C allele of the rs956115arker (Table 3). In the external validation cohort of
atients without DM in their steady state phase of dualntiplatelet therapy, C allele carriers of the rs956115 markerepresented 31.1% of the patient population. The ADP-nd collagen-induced platelet aggregation were 53.2 �6.5% and 41.0 � 19.7%, respectively. Although there wasgreater prevalence of patients with HPR among patientho were carriers of the C allele of the rs956115 marker,
his was not statistically significant (Table 3).RS-1 haplotypes and platelet reactivity. The IRS-1 geneaplotypes were inferred from the 7 tag SNPs (Onlineable 7). Haplotype association tests were conducted on therimary sample (208 individuals) in which there was noignificant association between any of these inferred haplo-ypes and the phenotypes investigated in this study (data nothown). Three haplotype LD blocks were identified in theenotyped sample for the IRS-1 gene region: block-1rs11683087-rs2251692), block-2 (rs1801278-rs1801123-s6725330), and block-3 (rs1896832-rs956115) (see Onlineable 8 for frequencies, Online Fig. 2 for pairwise LD
tructure, and Online Table 9 for LD measures). HaplotypeD block analyses showed no significant associations forlock-1 or block-2 with any of the phenotypes investigated.here was a significant association between haplotypes ofD block-3 (rs1896832-rs956115) and HPR (global p �.036); a trend was observed with ADP- and collagen-nduced aggregation above the upper quartiles (global p �.09). Haplotype rs1896832-A/rs956115-C showed a sig-ificant association with HPR (adjusted OR: 2.63; CI: 1.26o 5.48; p � 0.01) and ADP- and collagen-induced aggre-ation above the upper quartiles (adjusted OR: 2.36; CI:.02 to 5.47; p � 0.045).
ding to rs956115 Genotypes of the IRS-1 Gene
the insulin receptor substrate-1 (IRS-1) gene have a greater prevalence of highThe prevalence of carriers of the C allele increases across quartile (Q) distribu-
ccor
pe ofs) (A).
tcdwMc9C
are tesbetes m
35JACC Vol. 58, No. 1, 2011 Angiolillo et al.June 28, 2011:30–9 Gentics, Diabetes, and Outcomes
Platelet reactivity and clinical outcomes. Major adversecardiovascular events occurred in a total of 28 patients (15%)during the 24-month follow-up period. Major adversecardiovascular events were largely driven by ACS requiringhospital stay (n � 26; 93%); 2 patients experienced acardiovascular death and a nonfatal ischemic stroke. Therewere a total of 4 noncardiac deaths. Patients with HPR wereat significantly higher risk of MACE (28.0% vs. 10.9%;HR: 2.90, 95% CI: 1.38 to 6.11; p � 0.005) (Fig. 2). Asignificant association between HPR and MACE wasconfirmed in the multivariable analysis (adjusted HR: 3.10,95% CI: 1.47 to 6.52; p � 0.003). In patients with HPR, arend toward a higher risk of MACE was observed duringlopidogrel treatment (6.4% vs. 2.1%; p � 0.16), whereasifferences in outcomes were statistically significant after itsithdrawal (26.8% vs. 9.8%; p � 0.01). A higher risk ofACE was seen in patients in the upper quartile of
ollagen-induced aggregation (23.9% vs. 12.3%; HR: 2.13,5% CI: 1.00 to 4.55; p � 0.051; adjusted HR: 2.18, 95%I: 1.02 to 4.67; p � 0.044) and in patients with both
ADP- and collagen-induced aggregation above the upperquartiles (35.2% vs. 11.9%; HR: 3.49, 95% CI: 1.58 to 7.71;p � 0.002; adjusted HR: 3.71, 95% CI: 1.67 to 8.25;p � 0.001) (Fig. 2). The upper quartiles of ADP- andcollagen-induced platelet aggregation were found to be thebest predictors of MACE in the ROC analyses.IRS-1 genotypes and clinical outcomes. Carriers of the Callele of the rs956115 marker had a significantly higher riskof MACE (30.6% vs. 11.4%; HR: 2.88, 95% CI: 1.35 to6.14; p � 0.006) (Fig. 3). Carriers of the C allele had anonsignificant increase in MACE while receiving clopi-dogrel treatment (5.6% vs. 2.6%; p � 0.39). Major adverse
Prevalence of Patients With HPR Among Carrierof the C Allele of the rs956115 MarkerTable 3 Prevalence of Patients With HPR Aof the C Allele of the rs956115 Ma
Patients With HPR
Prim
Total(n � 187)
Carriers(n � 36)
After ADP 47 (25.1) 16 (44.4)
After collagen 47 (25.1) 14 (38.9)
After ADP and collagen 26 (13.9) 10 (27.8)
Validation Coho
Total(n � 52)
Carriers(n � 10)
After ADP 4 (7.7) 3 (30.0)
After collagen 9 (17.3) 6 (60.0)
After ADP and collagen 2 (3.8) 2 (20.0)
Validation Cohor
Total(n � 90)
Carriers(n � 28)
After ADP 23 (25.6) 10 (35.7)
After collagen 18 (20.0) 8 (28.6)
After ADP and collagen 14 (15.6) 6 (21.4)
Values are n (%). *The p values are calculated with the use of chi-squADP � adenosine-diphosphate; CI � confidence interval; DM � dia
cardiovascular events increased over time after clopidogrel
withdrawal in C allele carriers (31.2% vs. 9.9%; p � 0.005)(see Online Fig. 3 for landmark analysis). In the multivari-able analysis, the C allele of the rs956115 marker showed tobe an independent predictor of MACE both in the modelnot including (adjusted HR: 3.11, 95% CI: 1.45 to 6.68;p � 0.004) and in that including HPR (adjusted HR: 2.31,95% CI: 1.03 to 5.19; p � 0.04) as a covariate. There wereno differences in baseline demographic data and clinicalcharacteristics of patients with (n � 36) and without (n �151) the C allele of the rs956115 marker (Online Table 10).There was no interaction according to insulin usage onHPR (p for interaction � 0.58) and MACE (p for inter-action � 0.82) (Online Table 11). There was a significantassociation between haplotypes of LD block-3 (rs1896832-rs956115) and MACE (global p � 0.028). Haplotypers1896832-A/rs956115-C showed a significant associationwith MACE (adjusted OR: 3.0, 95% CI: 1.36 to 6.7; p �0.007).
Discussion
This is the first study to evaluate the impact of genesequence variations on antiplatelet drug effects and clinicaloutcomes in patients with DM. In particular, the results ofthe present study demonstrate that, in patients with type 2DM and stable coronary artery disease, gene sequencevariations of IRS-1—namely C allele carriers of thers956115 polymorphism (observed in approximately 20% ofpatients)—associate independently with a hyperreactiveplatelet phenotype and enhanced long-term cardiovascularrisk. These findings not only provide further insights onpharmacogenetic modulation of antiplatelet drug effects but
NoncarriersCarriers and Noncarriers
ohort
OR (95% CI)Noncarriers(n � 151) p Value*
31 (20.5) 0.006 3.10 (1.44–6.67)
33 (21.9) 0.058 2.28 (1.05–4.93)
16 (10.6) 0.01 3.25 (1.33–7.93)
atients With DM
Noncarriers(n � 42) p Value*
1 (2.4) 0.02 17.6 (1.59–193.88)
3 (7.1) 0.001 19.5 (3.47–109.57)
0 0.03 —
tients Without DM
Noncarriers(n � 62) p Value*
13 (21.0) 0.22 2.09 (0.78–5.61)
10 (16.1) 0.28 2.08 (0.72–6.02)
8 (12.9) 0.35 1.84 (0.57–5.92)
t or Fisher exact test as appropriate.ellitus; HPR � high platelet reactivity; OR � odds ratio.
s andmongrker
ary C
rt of P
t of Pa
also provide a genetic explanation as to why variable clinical
36 Angiolillo et al. JACC Vol. 58, No. 1, 2011Gentics, Diabetes, and Outcomes June 28, 2011:30–9
outcomes might occur within a population, such as thosewith DM homogeneous for baseline risk profile.
Variability in individual response to antiplatelet therapy isan emerging clinical entity (3,4). The mechanisms leadingto antiplatelet drug response variability are not fully estab-lished and are likely multifactorial (3,4). Pharmacogeneticshas recently emerged as a field that tries to explain thisphenomenon (10). Recent findings have shown genetictargets modulating pharmacokinetic profiles of clopidogrelthrough its metabolism by the cytochrome P450 (CYP)enzymatic system to have a major role on its pharmacody-namic effects (27–32). This might explain why recentstudies have shown that gene sequence variations ofCYP2C19 are associated with an increased risk of adverseevents in clopidogrel-treated patients (30–35). However,gene sequence variations of CYP2C19 contribute to onlyapproximately 12% of the interindividual response profile toclopidogrel (32), and these findings cannot be extrapolatedto patients with DM who have specific aberrations in theirplatelet function compared with patients without DM,leading to differences in pharmacodynamic profiles that areultimately determinants of thrombotic mediated processes(5–8). In fact, in vitro and ex vivo studies have shown thatreduced pharmacodynamic effects of antiplatelet agents inpatients with DM are attributed to upregulation of plateletsignaling pathways (5–9,12), suggesting the potential mod-ulating role of genetic determinants of “downstream” (e.g.,platelet membrane receptors) mediators of platelet reactiv-ity. Although glycemic control is known to be associatedwith platelet reactivity through various mechanisms, includ-ing glycation of platelet surface proteins (36), in our study
Figure 3 Association Between rs956115 Genotype Statusand Major Adverse Cardiac Events
Among 187 type 2 diabetic patients who were classified as carriers (GC andGG genotypes) or noncarriers (GG genotype) of the rs956115 genotype of theinsulin receptor substrate-1 (IRS-1) gene, the rate of major adverse ischemicevents (composite of CV death, ACS, or stroke) was 30.6% among carriers ascompared with 11.4% among noncarriers (HR: 2.88, 95% CI: 1.35 to 6.14).Abbreviations as in Figure 2.
Figure 2 Association Between Profiles of Platelet Reactivityand Major Adverse Cardiac Events
Among 187 type 2 diabetic patients who were classified as having highon-treatment platelet reactivity (HPR) after adenosine diphosphate (ADP) (A),collagen (B), or both (C) stimuli, the rate of major adverse ischemic events(composite of cardiovascular [CV] death, acute coronary syndrome [ACS], orstroke) was greater compared with those without (ADP: 28.0% vs. 10.9%, haz-ard ratio [HR]: 2.90, 95% confidence interval [CI]: 1.38 to 6.11; collagen:23.9% vs. 12.3%, HR: 2.13, 95% CI: 1.00 to 4.55; ADP�collagen: 35.2% vs.11.9%; HR: 3.49, 95% CI: 1.58 to 7.71).
this was not observed—likely because patients in our study
iplmrhr(cHspaTac(ahwopWtwepIhcrtn
37JACC Vol. 58, No. 1, 2011 Angiolillo et al.June 28, 2011:30–9 Gentics, Diabetes, and Outcomes
had good glycemic control and limited variability inHbA1C levels, as also shown in prior investigations (5,6).The impact of downstream genetic determinants are sup-ported by our study findings in which IRS-1 genotypes wereassociated with a hyperreactive platelet phenotype in pa-tients with type 2 DM but not in those without thismetabolic disorder. These pharmacodynamic effects wereconfirmed irrespective of whether patients were in the acutephase of treatment after a high loading dose regimen or inthe maintenance phase of dual antiplatelet therapy. A highloading dose of clopidogrel in patients undergoing PCIleads to enhanced platelet inhibitory effects with a broaderrange of variability compared with patients in their long-term maintenance of standard dosing (3), as also shown inthis study. More variable profiles of platelet reactivityenables better identification of whether there are specificfactors associated with poor response. This might explainwhy the ORs of having a hyperreactive platelet phenotypeamong carriers of the C allele of the rs956115 marker werehigher in the acute phase of therapy compared with themaintenance phase.
The present study further supports the prognostic impli-cations associated with a hyperreactive platelet phenotype(3,4). Of note, the magnitude of effect on clinical outcomes(approximately 3-fold increase in MACE) observed in ourstudy was of the same extent or even greater than thatobserved in recent CYP2C19 studies (31–35). It should beunderscored that the latter were performed in ACS patients,many undergoing PCI, in whom most events occurred earlyand survival curves paralleled over time, suggesting a prog-nostic role of CYP2C19 gene variants for early but not lateevents. In contrast, in our analysis, we studied stablepatients—in a period remote from when recurrent eventsmost commonly occur (6 to 9 months after PCI)—andshowed that survival curves diverge over time. Previousstudies, performed primarily in subjects without DM or inpre-DM states, have shown functional polymorphisms ofthe IRS-1 gene to modulate insulin sensitivity (13) as wellas to be a risk factor for coronary artery disease (14). In ourstudy population of patients with type 2 DM, however, wedid not find any interaction among the rs956115 marker,degree of insulin resistance, and platelet reactivity. Becausethe rs956115 polymorphism is located in the 5’ region of theIRS-1 gene, this does not affect amino acid coding and doesnot directly affect protein function. Therefore, our findingsmight be due to a linkage with other SNPs in exons(resulting in functional polymorphism) or in regulatoryregions (affecting the expression of IRS-1 gene). Thecomplexity of intraplatelet signaling and the potential forinterplay with other pathways that derives from IRS-1suggest that, although levels of insulin sensitivity remain acontributor to platelet function profiles, many other mech-anisms might be involved in determining a hyper-reactiveplatelet phenotype as a consequence of a dysfunction ofIRS-1–mediated signaling. Because the rs956115 C allele
was an independent predictor of clinical outcomes after
adjustment for potential confounders (with and withoutHPR as a covariate), other unknown reasons that are notentirely linked to HPR might be implied and warrantfurther investigation.
Although loss of responsiveness to insulin via IRS-1 hasshown to be associated with upregulation of P2Y12 signal-ng, it cannot be excluded that this might also affect otherlatelet signaling pathways, commonly upregulated in plate-
ets from patients with DM (12,37). In fact, IRS-1 is aajor tyrosine phosphorylated substrate for the insulin
eceptor acting as a multisite docking protein to several Srcomology 2 domains containing proteins, such as theegulatory subunits of phosphatidylinositol 3-kinasePI3K)—which are key in multiple platelet activation pro-esses (13). This is in line with the fact that patients withPR to ADP frequently have HPR to non-purinergic
timuli (i.e., collagen), indicative of an overall hyper-reactivelatelet phenotype, which might also be a better predictor ofdverse outcomes, as also suggested by this study (9,38,39).his might contribute to the elevated prevalence of reduced
spirin-induced antiplatelet effects when measured byyclooxygenase-1 nonspecific assays in patients with DM6,40–42). Furthermore, patients with DM presenting with
hyper-reactive platelet phenotype have been shown toave a marked increase in platelet reactivity after clopidogrelithdrawal (43). Overall, these findings might explain whyur long-term survival curves continue to diverge over time,articularly while patients were only taking aspirin therapy.
hether prolonging clopidogrel therapy in patients definedo be at higher risk on the basis of our laboratory findingsould have led to improved clinical outcomes cannot be
xtrapolated from this study. It might be hypothesized thatatients who are type 2 DM carriers of the C allele of theRS-1 rs956115 tag SNP, who our study demonstrated toave a hyper-reactive platelet phenotype and worse out-omes, might benefit from more potent antithromboticegimens. These might include high-dose clopidogrel (44),riple therapy (aspirin, clopidogrel, and cilostazol) (45), orovel and more potent P2Y12 receptor antagonists (46,47).
Among the latter, prasugrel has been shown to be associatedwith better clinical outcomes, particularly in patients withDM (48). However, atherothrombotic event rates continueto be high in patients with DM, which might be attributedto upregulation of other pivotal platelet signaling pathwaystriggering thrombosis, suggesting the need for antiplateletagents that are able to block these alternative pathways(46,49).
In summary, heterogeneous antiplatelet drug effects areobserved in type 2 DM patients, and patients with HPRhave a greater risk of recurrent events. The C allele of thers956115 polymorphism of IRS-1, observed in approxi-mately 20% of patients, is independently associated withHPR and enhanced long-term cardiovascular risk. Thesefindings might explain why in clinical practice, althoughtype 2 DM represents per se a high-risk cohort, some
patients have worse outcomes than others and might war-
1
1
1
1
1
1
1
1
1
1
2
38 Angiolillo et al. JACC Vol. 58, No. 1, 2011Gentics, Diabetes, and Outcomes June 28, 2011:30–9
rant more aggressive antithrombotic treatment. Our obser-vations provide further insights on how pharmacogeneticanalyses might identify patients with type 2 DM at differentcardiovascular risk, suggesting the need for personalizedtreatment strategies in these patients.Study limitations. Several cutoff values of HPR have beendefined in the published data, although these might varyaccording to the specific population under investigation ortiming from an acute event, among many other variables(50). Because this study selectively investigated a populationwith DM in a period remote from their PCI for which thereis limited data on cutoff values of HPR, in agreement withprior investigations, we considered a ROC analysis to definethe value with the highest sensitivity and highest specificityin our study population (50). Furthermore, it might beargued that, although marker rs956115 is in LD withrs1896832, only rs956115 showed a significant associationwith outcome measures. This can be explained by the degreeof LD between the 2 markers. In fact, although the C alleleof rs956115 is fully associated with the A allele ofrs1896832, the opposite is not true, because most of the Aalleles of marker rs1896832 are not associated with the Callele of marker rs956115 (approximately 88%). This isbecause alleles at 2 different markers have different frequen-cies and thus not the same as in the case of an absolute LD.Because in the association analysis of marker rs1896832only a proportion of A alleles has a different effect comparedwith the other G alleles (those in linkage with the C alleleof marker rs956115), a larger sample of individuals asreported by others would be required to detect a significanteffect of allele A (51). An independent validation would alsoallow a better estimation of allelic frequencies of the IRS-1gene, which in our study showed some differences, likelyattributable to the sample size, in patients with and withoutDM. Although the advantage of using a study populationwithout very strong LD between SNP to map causalvariants might be a good strategy to identify portions ofgenes implicated in the susceptibility of the phenotypes ofinterest, indeed sequencing of the entire gene and promoterregion is the definitive approach to identify all the importantsequence variants. Ultimately, in the logistic regression forstatistical adjustment, the potential for overfitting of themodel may not be excluded, because the selection of thevariables was based on clinical judgment by entering all thecovariates that might impact the degree of plateletaggregation.
Reprint requests and correspondence: Dr. Dominick J Angio-lillo, Division of Cardiology, University of Florida College ofMedicine-Jacksonville, 655 West 8th Street, Jacksonville, Florida32209. E-mail: [email protected].
REFERENCES
1. Anderson JL, Adams CD, Antman EM, et al. ACC/AHA 2007
guidelines for the management of patients with unstable angina/non–
ST-elevation myocardial infarction: a report of the American Collegeof Cardiology/American Heart Association Task Force on PracticeGuidelines J Am Coll Cardiol 2007;50:e1–157.
2. Kushner FG, Hand M, Smith SC Jr., et al. 2009 focused updates:ACC/AHA guidelines for the management of patients with ST-elevation myocardial infarction (updating the 2004 guideline and 2007focused update) and ACC/AHA/SCAI guidelines on percutaneouscoronary intervention (updating the 2005 guideline and 2007 focusedupdate). J Am Coll Cardiol 2009;54:2205–41.
3. Angiolillo DJ, Fernandez-Ortiz A, Bernardo E, et al. Variability inindividual responsiveness to clopidogrel: clinical implications, manage-ment, and future perspectives. J Am Coll Cardiol 2007;49:1505–16.
4. Angiolillo DJ. Variability in responsiveness to oral antiplatelet therapy.Am J Cardiol 2009;103:27A–34A.
5. Angiolillo DJ, Fernandez-Ortiz A, Bernardo E, et al. Platelet functionprofiles in patients with type 2 diabetes and coronary artery disease oncombined aspirin and clopidogrel treatment. Diabetes 2005;54:2430–5.
6. Angiolillo DJ, Bernardo E, Ramirez C, et al. Insulin therapy isassociated with platelet dysfunction in patients with type 2 diabetesmellitus on dual oral antiplatelet treatment. J Am Coll Cardiol2006;48:298–304.
7. Geisler T, Anders N, Paterok M, et al. Platelet response to clopidogrelis attenuated in diabetic patients undergoing coronary stent implanta-tion. Diabetes Care 2007;30:372–4.
8. Angiolillo DJ. Antiplatelet therapy in diabetes: efficacy and limitationsof current treatment strategies and future directions. Diabetes Care2009;32:531–40.
9. Angiolillo DJ, Bernardo E, Sabaté M, et al. Impact of plateletreactivity on cardiovascular outcomes in patients with type 2 diabetesmellitus and coronary artery disease. J Am Coll Cardiol 2007;50:1541–7.
0. Marin F, Gonzalez-Conejero R, Capranzano P, Bass TA, Roldan V,Angiolillo DJ. Pharmacogenetics in cardiovascular antithrombotictherapy. J Am Coll Cardiol 2009;54:1041–57.
1. Ferreira IA, Eybrechts KL, Mocking AI, Kroner C, Akkerman JW.IRS-1 mediates inhibition of Ca2� mobilization by insulin via theinhibitory G-protein Gi. J Biol Chem 2004;279:3254–64.
2. Ferreira IA, Mocking AI, Feijge MA, et al. Platelet inhibition byinsulin is absent in type 2 diabetes mellitus. Arterioscler Thromb VascBiol 2006;26:417–22.
3. Sesti G, Federici M, Hribal ML, Lauro D, Sbraccia P, Lauro R.Defects of the insulin receptor substrate (IRS) system in humanmetabolic disorders. FASEB J 2001;15:2099–111.
4. Baroni MG, D’Andrea MP, Montali A, et al. A common mutation ofthe insulin receptor substrate-1 gene is a risk factor for coronary arterydisease. Arterioscler Thromb Vasc Biol 1999;19:2975–80.
5. Alberti KG, Zimmet PZ, for the WHO Consultation. Definition,diagnosis and classification of diabetes mellitus and its complications.Part 1: Diagnosis and classification of diabetes mellitus. Provisionalreport of a WHO consultation. Diabet Med 1998;15:539–53.
6. Cannon CP, Battler A, Brindis RG, et al., American College ofCardiology Task Force on Clinical Data Standards (Acute CoronarySyndromes Writing Committee). American College of Cardiology keydata elements and definitions for measuring the clinical managementand outcomes of patients with acute coronary syndromes. J Am CollCardiol 2001;38:2114–30.
7. Cuisset T, Frere C, Quilici J, et al. High post-treatment plateletreactivity identified low-responders to dual antiplatelet therapy atincreased risk of recurrent cardiovascular events after stenting for acutecoronary syndrome. J Thromb Haemost 2006;4:542–9.
8. Patti G, Nusca A, Mangiacapra F, Gatto L, D’Ambrosio A, DiSciascio G. Point-of-care measurement of clopidogrel responsivenesspredicts clinical outcome in patients undergoing percutaneous coro-nary intervention results of the ARMYDA-PRO (Antiplatelet therapyfor Reduction of MYocardial Damage during Angioplasty-PlateletReactivity Predicts Outcome) study. J Am Coll Cardiol 2008;52:1128–33.
9. Halperin E, Kimmel G, Shamir R. Tag SNP selection in genotypedata for maximizing SNP prediction accuracy. Bioinformatics 2005;21:i195–203.
0. Davidovich O, Kimmel G, Shamir R. GEVALT: an integrated
software tool for genotype analysis. BMC Bioinformatics 2007;8:36.
39JACC Vol. 58, No. 1, 2011 Angiolillo et al.June 28, 2011:30–9 Gentics, Diabetes, and Outcomes
21. GEnotype Visualization and ALgorithmic Tool. Available at: http://acgt.cs.tau.ac.il/gevalt/. Accessed April 2011.
2. Lewontin RC. On measures of gametic disequilibrium. Genetics1988;120:849–52.
3. Barrett JC, Fry B, Maller J, Daly MJ. Haploview: analysis andvisualization of LD and haplotype maps. Bioinformatics 2005;21:263–5.
4. Breet NJ, van Werkum JW, Bouman HJ, et al. Comparison of plateletfunction tests in predicting clinical outcome in patients undergoingcoronary stent implantation. JAMA 2010;303:754–62.
5. Lake S, Lyon H, Silverman E, Weiss S, Laird N, Schaid D.Estimation and tests of haplotype-environment interaction whenlinkage phase is ambiguous. Human Hered 2003;55:56–65.
6. Ridker PM, Buring JE, Rifai N, Cook NR. Development andvalidation of improved algorithms for the assessment of global cardio-vascular risk in women: the Reynolds Risk Score. JAMA 2007;297:611–9.
7. Angiolillo DJ, Fernandez-Ortiz A, Bernardo E, et al. Contribution ofgene sequence variations of the hepatic cytochrome P450 3A4 enzymeto variability in individual responsiveness to clopidogrel. ArteriosclerThromb Vasc Biol 2006;26:1895–900.
8. Hulot JS, Bura A, Villard E, et al. Cytochrome P450 2C19 loss-of-function polymorphism is a major determinant of clopidogrel respon-siveness in healthy subjects. Blood 2006;108:2244–7.
9. Brandt JT, Close SL, Iturria SJ, et al. Common polymorphisms ofCYP2C19 and CYP2C9 affect the pharmacokinetic and pharmaco-dynamic response to clopidogrel but not prasugrel. J Thromb Haemost2007;5:2429–36.
0. Trenk D, Hochholzer W, Fromm MF, et al. Cytochrome P450 2C19681G�A polymorphism and high on-clopidogrel platelet reactivityassociated with adverse 1-year clinical outcome of elective percutane-ous coronary intervention with drug-eluting or bare-metal stents. J AmColl Cardiol 2008;51:1925–34.
1. Mega JL, Close SL, Wiviott SD, et al. Cytochrome P450 poly-morphisms and response to clopidogrel. N Engl J Med 2009;360:354 – 62.
2. Shuldiner AR, O’Connell JR, Bliden KP, et al. Association ofcytochrome P450 2C19 genotype with the antiplatelet effect andclinical efficacy of clopidogrel therapy. JAMA 2009;302:849–57.
3. Simon T, Verstuyft C, Mary-Krause M, et al. Genetic determinants ofresponse to clopidogrel and cardiovascular events. N Engl J Med2009;360:363–75.
4. Collet JP, Hulot JS, Pena A, et al. Cytochrome P450 2C19 polymor-phism in young patients treated with clopidogrel after myocardialinfarction: a cohort study. Lancet 2009;373:309–17.
5. Wallentin L, James S, Storey RF, et al., PLATO investigators. Effectof CYP2C19 and ABCB1 single nucleotide polymorphisms on out-comes of treatment with ticagrelor versus clopidogrel for acute coro-nary syndromes: a genetic substudy of the PLATO trial. Lancet2010;376:1320–8.
6. Osende JI, Badimon JJ, Fuster V, et al. Blood thrombogenicity in type2 diabetes mellitus patients is associated with glycemic control. J AmColl Cardiol 2001;38:1307–12.
7. Ueno M, Ferreiro JL, Tomasello SD, et al. Functional profile of theplatelet P2Y12 receptor signalling pathway in patients with type 2diabetes mellitus and coronary artery disease. Thromb Haemost2011;105:730–2.
detects platelet hyperreactivity in healthy individuals. Blood 2005;106:2723–9.
39. Gori AM, Marcucci R, Paniccia R, et al. Thrombotic events in highrisk patients are predicted by evaluating different pathways of plateletfunction. Thromb Haemost 2008;100:1136–45.
40. Angiolillo DJ, Fernandez-Ortiz A, Bernardo E, et al. Influence ofaspirin resistance on platelet function profiles in patients on long-termaspirin and clopidogrel after percutaneous coronary intervention. Am JCardiol 2006;97:38–43.
41. DiChiara J, Bliden KP, Tantry US, et al. The effect of aspirin dosingon platelet function in diabetic and nondiabetic patients: an analysisfrom the aspirin-induced platelet effect (ASPECT) study. Diabetes2007;56:3014–9.
42. Capodanno D, Patel A, Dharmashankar K, et al. Pharmacodynamiceffects of different aspirin dosing regimens in type 2 diabetes mellituspatients with coronary artery disease. Circ Cardiovasc Interv 2011;4:180–7.
43. Angiolillo DJ, Fernandez-Ortiz A, Bernardo E, et al. Clopidogrelwithdrawal is associated with proinflammatory and prothromboticeffects in patients with diabetes and coronary artery disease. Diabetes2006;55:780–4.
44. Angiolillo DJ, Shoemaker SB, Desai B, et al. Randomized comparisonof a high clopidogrel maintenance dose in patients with diabetesmellitus and coronary artery disease: results of the Optimizing Anti-platelet Therapy in Diabetes Mellitus (OPTIMUS) study. Circulation2007;115:708–16.
45. Angiolillo DJ, Capranzano P, Goto S, et al. A randomized studyassessing the impact of cilostazol on platelet function profiles inpatients with diabetes mellitus and coronary artery disease on dualantiplatelet therapy: results of the OPTIMUS-2 study. Eur Heart J2008;29:2202–11.
46. Angiolillo DJ, Bhatt DL, Gurbel PA, Jennings LK. Advances inantiplatelet therapy: agents in clinical development. Am J Cardiol2009;103:40A–51A.
47. Angiolillo DJ, Badimon JJ, Saucedo JF, et al. A pharmacodynamiccomparison of prasugrel vs. high-dose clopidogrel in patients with type2 diabetes mellitus and coronary artery disease: results of the Opti-mizing anti-Platelet Therapy In diabetes MellitUS (OPTIMUS)-3Trial. Eur Heart J 2011;32:838–46.
48. Wiviott SD, Braunwald E, Angiolillo DJ, et al., TRITON–TIMI 38Investigators. Greater clinical benefit of more intensive oral antiplate-let therapy with prasugrel in patients with diabetes mellitus in the trialto assess improvement in therapeutic outcomes by optimizing plateletinhibition with prasugrel–Thrombolysis in Myocardial Infarction 38.Circulation 2008;118:1626–36.
49. Ferreiro JL, Angiolillo DJ. Diabetes and antiplatelet therapy in acutecoronary syndrome. Circulation 2011;123:798–813.
50. Bonello L, Tantry US, Marcucci R, et al., Working Group on HighOn-Treatment Platelet Reactivity. Consensus and future directions onthe definition of high on-treatment platelet reactivity to ADP. J AmColl Cardiol 2010;56:919–33.
51. Florez JC, Sjögren M, Agapakis CM, et al. Association testing ofcommon variants in the insulin receptor substrate-1 gene (IRS1) withtype 2 diabetes. Diabetologia 2007;50:1209–17.
Key Words: antiplatelet therapy y diabetes mellitus y gene y plateletfunction.
